wylab/llama-swap-rocm
1
0
Fork 0
forked from wylab/llama-swap
Code Pull Requests Actions Packages Activity

docker: make CMAKE_CUDA_ARCHITECTURES configurable via build arg (#625)

Browse Source 
Expose CMAKE_CUDA_ARCHITECTURES as a Docker build ARG so users can
customize CUDA architectures via --build-arg without editing the
Dockerfile.

- convert hardcoded ENV to ARG with default, feeding into ENV
- replace silent fallback defaults (:-) in scripts with :? guards
  to fail fast if the env var is missing
- add usage example to Dockerfile header

Follow up to: #624

https://claude.ai/code/session_01EWiUe7jNABX7Uz95dUGJqK

Co-authored-by: Claude <noreply@anthropic.com>
This commit is contained in:
Benson Wong
2026-04-04 08:49:59 +08:00
committed by GitHub
parent 1dd1aadf93
commit a185efe37e
5 changed files with 7 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docker/unified/Dockerfile
+3 -1
View File
@@ -4,6 +4,7 @@
# Usage:
# docker buildx build --build-arg BACKEND=cuda -t llama-swap:unified-cuda .
# docker buildx build --build-arg BACKEND=vulkan -t llama-swap:unified-vulkan .
# docker buildx build --build-arg BACKEND=cuda --build-arg CMAKE_CUDA_ARCHITECTURES="86;89" -t llama-swap:unified-cuda .
#
# Each project has its own install script that handles cloning, building,
# and installing binaries. Build stages are independent for cache efficiency.
@@ -14,8 +15,9 @@ ARG BACKEND=cuda
FROM nvidia/cuda:12.9.1-devel-ubuntu24.04 AS builder-base-cuda
ARG CMAKE_CUDA_ARCHITECTURES="60;61;75;86;89"
ENV DEBIAN_FRONTEND=noninteractive
ENV CMAKE_CUDA_ARCHITECTURES="60;61;75;86;89"
ENV CMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES}
ENV CCACHE_DIR=/ccache
ENV CCACHE_MAXSIZE=2G
ENV PATH="/usr/lib/ccache:${PATH}"
docker/unified/install-ik-llama.sh
+1 -1
View File
@@ -26,7 +26,7 @@ CMAKE_FLAGS=(
-DCMAKE_C_COMPILER_LAUNCHER=ccache
-DCMAKE_CXX_COMPILER_LAUNCHER=ccache
-DGGML_CUDA=ON
"-DCMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES:-60;61;75;86;89}"
"-DCMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES:?CMAKE_CUDA_ARCHITECTURES must be set}"
"-DCMAKE_CUDA_FLAGS=-allow-unsupported-compiler"
"-DCMAKE_EXE_LINKER_FLAGS=-Wl,-rpath-link,/usr/local/cuda/lib64/stubs -lcuda -Wl,--allow-shlib-undefined"
)
docker/unified/install-llama.sh
+1 -1
View File
@@ -33,7 +33,7 @@ if [ "$BACKEND" = "cuda" ]; then
CMAKE_FLAGS+=(
-DGGML_CUDA=ON
-DGGML_VULKAN=OFF
"-DCMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES:-60;61;75;86;89}"
"-DCMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES:?CMAKE_CUDA_ARCHITECTURES must be set}"
"-DCMAKE_CUDA_FLAGS=-allow-unsupported-compiler"
"-DCMAKE_EXE_LINKER_FLAGS=-Wl,-rpath-link,/usr/local/cuda/lib64/stubs -lcuda"
)
docker/unified/install-sd.sh
+1 -1
View File
@@ -33,7 +33,7 @@ if [ "$BACKEND" = "cuda" ]; then
CMAKE_FLAGS+=(
-DGGML_CUDA=ON
-DGGML_VULKAN=OFF
"-DCMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES:-60;61;75;86;89}"
"-DCMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES:?CMAKE_CUDA_ARCHITECTURES must be set}"
"-DCMAKE_CUDA_FLAGS=-allow-unsupported-compiler"
"-DCMAKE_EXE_LINKER_FLAGS=-Wl,-rpath-link,/usr/local/cuda/lib64/stubs -lcuda"
"-DCMAKE_SHARED_LINKER_FLAGS=-Wl,-rpath-link,/usr/local/cuda/lib64/stubs -lcuda"
docker/unified/install-whisper.sh
+1 -1
View File
@@ -31,7 +31,7 @@ if [ "$BACKEND" = "cuda" ]; then
CMAKE_FLAGS+=(
-DGGML_CUDA=ON
-DGGML_VULKAN=OFF
"-DCMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES:-60;61;75;86;89}"
"-DCMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES:?CMAKE_CUDA_ARCHITECTURES must be set}"
"-DCMAKE_CUDA_FLAGS=-allow-unsupported-compiler"
"-DCMAKE_EXE_LINKER_FLAGS=-Wl,-rpath-link,/usr/local/cuda/lib64/stubs -lcuda"
"-DCMAKE_SHARED_LINKER_FLAGS=-Wl,-rpath-link,/usr/local/cuda/lib64/stubs -lcuda"